Uber Developers

Wearables & Voice

Privileged and Confidential This endpoint design has been confidentially shared with you. It is still under development and is subject to change without notice. Please do not share this document or API endpoint details with anyone who is not authorized to have access. For more information read about scopes.

¶ Overview

Uber Rides MCP works with any AI agent, including assistants on smart glasses, watches and earbuds. The tools are the same — you just show the details by display or voice.

Ride estimates are available today. Voice confirmation and booking are coming soon.

¶ How it works on a wearable

Step	What to do
Ask	Turn the spoken request into a pickup and destination, then call `get_estimates_between_two_locations`.
Show	On a display, show the ride, price and pickup time. On audio-only devices, read them back in one sentence.
Confirm	Coming soon — a spoken “book it” will confirm the booking.
Book	Coming soon — `book_ride` will book the ride and announce it.

¶ Keep it short

Riders are often on the move. Keep prompts to one sentence and use place names rather than full addresses.

❌ Too long	✓ Better
“I found an UberX available for pickup at your current location at 123 Main Street. The estimated fare is $25.34 and a driver will arrive in approximately 4 minutes.”	“UberX in four minutes, twenty-five thirty-four.”

¶ Conversation pattern

Voice:  "How much is a ride to the airport?"

        → get_estimates_between_two_locations(destination: "SEA airport")

Lens:   UberX · 6 min · $41.20

Voice:  "UberX in six minutes, forty-one twenty."

¶ Audio-only devices

When there is no display (earbuds, voice assistants), read the estimate back as a single sentence:

"UberX to the airport — six minutes, forty-one twenty."

¶ Supported surfaces

Smart glasses (display + voice)
Smartwatches (small display + voice)
Earbuds / headphones (audio only)
Voice assistants (audio only)
Any wearable assistant that speaks MCP